Key Influencer visual - Behind the scene

[Anonymous]

Hi, I am trying to understand how the Key Influencer works behind the scene. In the documentation: https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-influencers#next-steps, it mentioned that the visual will also consider how many data points a factor has, using a Wald test of p-value 0.05 to determine the threshold. Does anyone know what this means?

[v-yulgu-msft]

Hi @Anonymous ,

Below links might give you a summary understanding about Wald test and P value.

Wald Test: Definition, Examples, Running the Test

WHAT A P-VALUE TELLS YOU ABOUT STATISTICAL DATA

Best regards,

Yuliana Gu

[Anonymous]

Hi @v-yulgu-msft ,

Thank you for answering. I have a sample data which shows the performance of different groups of student in an exam. Using the key influencer visual, it shows some bizzard result. For example, it treated a group with 9 students as "influential", while other groups which have 70+ students and higher number in that cateogry than average are not count as influential. In the documentation, it claims that this visual would not consider categories with small sample size, but from my example it doesn't seem to be the case. Can anyone explain why this is?

Here are some snapshot of the problem. I tried to upload the pbix file but can't find a way to do it.

Many Thanks,

JC